French genomic experience: genomics for all ruminant species Eric Venot, Anne Barbat, Didier Boichard, Pascal Croiseau, Vincent Ducrocq, Rachel Lefebvre, Florence Phocas, Marie Pierre Sanchez, Thierry Tribout, Aurelie Vinet, et al. To cite this version: Eric Venot, Anne Barbat, Didier Boichard, Pascal Croiseau, Vincent Ducrocq, et al.. French genomic experience: genomics for all ruminant species. 2016 Interbull Meeting, Oct 2016, Puerto Varas, Chile. hal-02743221 HAL Id: hal-02743221 https://hal.inrae.fr/hal-02743221 Submitted on 3 Jun 2020 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. French genomic experience: genomics for all ruminant species E. Venot1, A. Barbat1, D. Boichard1, P. Croiseau1, V. Ducrocq1, R. Lefebvre1, F. Phocas1, M-P Sanchez1, T. Tribout1, A. Vinet1, MN Fouilloux2, A. Govignon-Gion2, A. Launay2, J. Promp2, M. Barbat3, A. Baur3, C. Hoze3, S. Fritz3, R. Saintilan3, C. Carillier4, H. Larroque4, A. Legarra4, I. Palhière4, C. Robert-Granié4, R. Rupp4, F. Tortereau4, JM. Astruc5, V. Clement5, V. Loywyck5, P. Boulesteix6, S. Mattalia2 1 GABI, INRA, AgroParisTech, Université Paris-Saclay, 78350 Jouy-en-Josas, France [email protected] (Corresponding Author) 2 Institut de l’Elevage, GABI, domaine de Vilvert, 78350 Jouy-en-Josas, France 3 Allice, GABI, domaine de Vilvert, 78350 Jouy-en-Josas, France 4 GenPhySE, Université de Toulouse, INRA, INPT, ENVT, 31326 Castanet-Tolosan, France 5 Institut de l’Elevage, 31326 Castanet Tolosan, France 6 Institut de l’Elevage, 87060 Limoges, France Abstract Genomic selection was implemented in France in 2009 for 3 French dairy cattle breeds and was then extended to 3 beef and 5 extra dairy cattle breeds. It was also implemented in the Lacaune dairy sheep breed in 2015, and will start in 2017 for the 2 main goat breeds and 3 other dairy sheep breeds. Major genes information will be used in meat sheep in the near future. For each species and breed, the genomic methodology has been adapted to account for the specificities regarding genomic evaluation: size of the reference population, linkage disequilibrium, multi-breed evaluations, addition of foreign genetic and genomic information, cost and benefits compared across species. A strong partnership between Research and Development organizations and industry made it possible to implement this technology in large populations as well as in small ones. This was facilitated through sharing of technological expertise and fast innovation transfer down to farm level. Version preprint Genomic tools allow selection for new traits that can be included in breeding goals. Use of genomic information is not limited to genomic evaluation: other new services of interest for breeders were developed, such as the determination of the animal status for specific major genes or genetic defects. Keywords: genomic selection, ruminant, dairy, meat Introduction Considering its size (550,000 km2), France presents a large landscape diversity and has managed to maintain a large panel of ruminant species and breeds adapted to these landscape specificities: 3 national dairy cattle breeds (Holstein (HOL), Montbéliarde (MON), Normande (NOR)) and 5 regional ones are nowadays under selection in France. The same can be found in dairy sheep mainly represented by the Lacaune breed (used for production of Roquefort cheese), followed by Manech (Red Faced and Black Faced), Basco-Béarnaise and Corsican breeds. The number of goat breeds is also large with 2 major breeds: Saanen and Alpine. Meat breeds are also well represented in France with in cattle, 3 national (Charolaise (CHA), Limousine (LIM) and Blonde d’Aquitaine (BLA)) and 6 regional (Aubrac (AUB), Bazadaise (BAZ), Gasconne (GAS), Parthenaise (PAR), Rouge des prés (ROU) and Salers (SAL)) beef breeds. France is also one of the European leaders in sheep meat production relying on more than 20 specialized or hardy breeds. France has managed to implement or will soon implement genomic selection for a large number of these breeds, thanks to a collaborative organization at national level. Comment citer ce document : Venot, E. (Auteur de correspondance), Barbat, A., Boichard, D., Croiseau, P., Ducrocq, V., Lefebvre, R., Phocas, F., Sanchez, M. P., Tribout, T., Vinet, A., Fouilloux, M.-N., Govignon Gion, The French genetic organization The French genetic organization was originally built through the 1966 Livestock Law. This legislation defined the general framework for animal breeding operations (animal identification, performance recording, breed association, semen production, semen distribution, extension). It also designated the French National Institute for Agricultural Research (INRA) as the center operating the central database gathering all pedigree and phenotype data and running genetic evaluations and the French Livestock Institute (Idele) for EBV delivery. This legislation was updated in 2006 by suppressing most monopolistic situations but confirming the roles of INRA and Idele and creating an inter-professional association for genetic improvement of all ruminants (France Génétique Elevage – FGE), in charge of the coordination of the overall national system for all ruminants. FGE represents France at the International Committee of Animal Recording (ICAR). A common R&D effort is carried out by two joint private-public structures (UMT 3G and UMT GGPR) bringing together INRA, Idele and the French Federation of breeding companies (Allice). The dairy cattle pioneers: 20 years of collective investments An ideal situation Developments of genomic evaluation first focused on the 3 national dairy cattle breeds (HOL, MON, NOR) because of their favorable characteristics: a large number of progeny tested bulls, a high proportion of AI, the high cost of progeny test, a reduced genetic population size, a relatively low genotyping cost compared with the value of breeding animals… Moreover, the major breed (HOL) is used internationally and participates to Interbull international evaluations. This international status helped to build up a consortium with other countries to exchange genotypes and therefore increase reference population size: the Eurogenomics consortium gathered 7 organizations from 8 countries (initially – in 2009- Denmark, Finland, France, Germany, the Netherlands, the Flanders part of Belgium and Sweden, and later Spain and Poland). A similar international initiative for genotype exchange took also place later in the Brown Swiss breed through the Intergenomics consortium. Reference population sizes and expected reliability for candidates are presented in Table 1. Version preprint Table 1. Size of the male and female reference populations in dairy cattle breeds in 2016 and range of (trait dependent) expected reliability for young genotyped animals. Breed Number of Bulls Number of Cows Expected reliability for young Candidates HOLSTEIN & RED 33 000 [0.55-0.70] BROWN SWISS 6 000 [0.45-0.70] MONTBELIARDE 2 800 31 000 [0.55-0.70] NORMANDE 2 400 16 000 [0.55-0.65] ABONDANCE 350 1 900 [0.35-0.55] TARENTAISE 300 1 300 [0.30-0.50] VOSGIENNE 60 1 100 [0.20-0.50] National genomic selection of dairy cattle The current approach (an extension of Boichard et al, 2012) consists in a genomic evaluation which combines the advantages of QTL-detection and genomic selection (GS): for each trait, a QTL detection is first applied using a Bayesian method (Bayes C). Haplotypes of 4 markers are then defined for each of the 3000 main QTL. It is expected that haplotypes trace better the transmission of QTL between generations. A residual polygenic component is added to estimate the sum of the remaining (tiny) QTL, through a genomic relationship matrix based on about 8000 SNP of the low density chip. In practice, a completely equivalent model is used where haplotype and SNP effects are estimated together. Comment citer ce document : Venot, E. (Auteur de correspondance), Barbat, A., Boichard, D., Croiseau, P., Ducrocq, V., Lefebvre, R., Phocas, F., Sanchez, M. P., Tribout, T., Vinet, A., Fouilloux, M.-N., Govignon Gion, Official genomic selection started in 2009 for HOL, MON and NOR breeds, in 2013 for BSW and in 2016 for regional breeds (ABO, TAR, VOS). Genomic evaluations of females became official in 2011. National evaluations are run 6 times per year (including haplotype and SNP effects re- estimation) and weekly for all animals with new genotype (without haplotype and SNP effects re- estimation). Since 2013, most genotypes are obtained with the EuroG10k chip and imputed to the 50k chip. Nevertheless, all AI bulls are (re)genotyped with the 50k chip to maintain a high imputation accuracy. Practical implications Genomic selection was quickly implemented in the 3 national dairy cattle breeds and quickly had large impacts on selection schemes of these breeds. No new young bull entered progeny test after 2010. With GS, GEBV of any genotyped animal are available at the latest 3 months after birth with already high reliability for all kinds of traits (between 0.6 and 0.7 for production or functional traits). Bulls are selected and used on a large scale as soon as they reach sexual maturity. Regular genetic evaluations are also performed, so EBV of these bulls based on progeny performances are still available (and compared with initial GEBV) but with a much higher reliability, because of the much larger number of progeny per bull. Dairy breeders quickly adopted this technology: since 2011, the number of genotyped calves increased by more than 25% per year, reaching 140,000 in 2016. Breeders also quickly changed their practices and now mainly use young genomic bulls: in 2015, 68%, 67% and 90% of the AI corresponded to young bulls with no daughter information yet, in HOL, MON, and NOR, respectively.